Neuromuscular diseases are degenerative and, thus far, incurable disorders that lead to large muscle wasting. They result in constant deterioration of activities of daily living and in particular of ambulation. Some common types include Duchenne muscular dystrophy, Charcot-Marie-Tooth disease, polymyositis and amyotrophic lateral sclerosis. While these diseases individually have a low rate of occurrence and are mostly unknown to most people, collectively they affect a significant part of the population. About 1 person in 2000 suffer from neuromuscular diseases, which means an approximate total of 370’000 people over the European continent. Recent technology breakthroughs have made possible the realization of advanced powered orthotics, which are commonly called exoskeletons. The most advanced devices have successfully been able to support patients in walking despite a debilitating condition such as complete spinal cord injury. Such technology could be ideal for people with mid-stage neuromuscular diseases as it provides more mobility and independence. This work investigates the definitions and requirements that would need to be fulfilled for any proposed orthotic device to assist people living with neuromuscular diseases. To define the needs of patients with neuromuscular disease, a large literature review is conducted on gait compensation patterns. The research also includes the data collection of experimental gait measurements from fourteen people with heterogeneous neuromuscular diseases. Conclusions show that orthotics for people with neuromuscular diseases require tunable assistance at each joint and a collaborative control strategy in order to let the user control motion. Eventually, most people may not be able to use crutches. A full lower limb exoskeleton, AUTONOMYO, is designed, realized and evaluated. A particular attention is put on the optimization of the actuator and transmission units. In order to reduce the effects of inertia and weight of those units, a design is explored with actuation remotely located from the joints. The transmission is realized by custom cable wire and pulley systems, combined with standard planetary gears. The dynamics of different coupling between the hip and the knee flexion/extension joints are explored, and their benefits and tradeoffs analyzed. A novel control strategy based on a finite-state active impedance model is designed and implemented on the AUTONOMYO device. The controller consists of three states of different active impedances mimicking a visco-elastic behavior. The switching condition between states is uniquely based on the hip flexion velocity to detect the user intent. The performance of the strategy regarding the detection of intention and the modulation of the assistance is evaluated on a test bench and in real conditions with healthy pilots and with a person with limb girdle muscular dystrophy. The preliminary results are promising since all pilots (including the one with muscular dystrophy) are able to initiate and terminate assisted walking on demand. They are all able both to walk with a good stride rate and to reach moderate velocities. Healthy pilots are able to ambulate alone with the exoskeleton, while the pilot with muscular dystrophy requires human assistance for the management of balance.